{
    "epoch": 1.0,
    "eval_logits/chosen": -2.841262102127075,
    "eval_logits/rejected": -2.8343887329101562,
    "eval_logps/chosen": -281.85919189453125,
    "eval_logps/rejected": -262.4202880859375,
    "eval_loss": 0.6618225574493408,
    "eval_pred_label": 13352.98046875,
    "eval_rewards/accuracies": 0.6150000095367432,
    "eval_rewards/chosen": 0.011613711714744568,
    "eval_rewards/margins": 0.02489962987601757,
    "eval_rewards/rejected": -0.013285920023918152,
    "eval_runtime": 834.3056,
    "eval_samples": 2000,
    "eval_samples_per_second": 2.397,
    "eval_steps_per_second": 0.3,
    "eval_use_label": 18709.01953125,
    "train_loss": 0.6705795382954063,
    "train_runtime": 45840.6595,
    "train_samples": 61135,
    "train_samples_per_second": 1.334,
    "train_steps_per_second": 0.021
}